Binary rhythm and syllable weight in
Portuguese
Guilherme D. Garcia
ball state university
guilhermegarcia.github.io
95th LSA Meeting
Jan 2021
Garcia Binary rhythm and syllable weight in Portuguese 2021
Intro
Secondary stress typically not a�ected by weight (Gordon 2006)
Finnish:
Primary stress is word initial
Secondary stress a�ected by weight1
(Lehiste 1965)
+ Light syllables skipped if followed by heavy syllable
1
Also see Pater (2000) for English.
Garcia Binary rhythm and syllable weight in Portuguese 1 of 21
Finnish
a. ka.las.te.let ‘you are fishing’
b. ka.las.te.lem.me ‘we are fishing’ skipping with H
c. fı.lo.so.fıs.sa ‘philosopher-ine’ no skipping with H
d. a.te.ri.a.na ‘meal-ess’ skipping with L
e. pro.fes.so.rıs.sa ∼ pro.fes.so.ris.sa ‘professor-ine’ variation
Sonority & clash avoidance play a role in variation (Anttila 2010)
+ Lexical⇔ post-lexical planes (Dresher and van der Hulst 1998)
Garcia Binary rhythm and syllable weight in Portuguese 2 of 21
Portuguese
Primary stress in non-verbs
Trisyllabic domain: σσσ]
Primary stress a�ected by weight (Wetzels 2007; Garcia 2017)
Regular stress:
XXH→ XXH jornal, papel ‘newspaper’, ‘paper’
XXL→ XXL borboleta, cavalo ‘butterfly’, ‘horse’
Irregular stress
XXL abacaxı ‘pineapple’
XXH nıvel ‘level’
XXX patetico ‘pathetic’
Garcia Binary rhythm and syllable weight in Portuguese 3 of 21
Portuguese
Secondary stress in non-verbs
Secondary stress weight-insensitive (Collischonn 1994, p. 45)
Stress every other pre-tonic syllable R-L
a. ınternacional ‘international’
borboleta ‘butterfly’
paralelepıpedo ‘block paving’
Garcia Binary rhythm and syllable weight in Portuguese 4 of 21
Portuguese
Secondary stress in non-verbs
Variation: odd number of pre-tonic syllables
b. cacofonıa ∼ cacofonıa ‘cacophony’
melhoramento ∼ melhoramento ‘improvement’
+ Either initial or peninitial secondary stress
Could weight a�ect variation in secondary stress?
Garcia Binary rhythm and syllable weight in Portuguese 5 of 21
Methods
Auditory judgment task in Praat (Boersma and Weenink 2020)
Nonce words in Portuguese (n = 120)
Native speakers of Brazilian Portuguese (n = 20)
stimuli
5-syl
LHLHLHLLHLLLLHL
4-syl
(controls)
LHHLHLHLLLHL
Minimal pairs: initial vs. peninitial secondary stress
argadorıste ∼ argadorıste
morilante ∼ morılante (control)
Garcia Binary rhythm and syllable weight in Portuguese 6 of 21
Methods
Stimuli
+ Correlate for stress in Portuguese: duration (Major 1985; Moraes 2003)
Secondary stress→ harder to capture:
Some (most?) studies point to duration (e.g., Moraes 2003)
Some to intensity (Fernandes-Svartman et al. 2008)
Similar situation for Spanish (Hualde and Nadeau 2014)
Garcia Binary rhythm and syllable weight in Portuguese 7 of 21
Methods
Duration in stimuli: 4σ (controls)
+ Durational pattern for secondary stress in stimuli
Recorded by native speaker with phonetic training
Initial
100ms
300ms
500ms
1 2 3 4
Syllable
Peninitial
100ms
300ms
500ms
1 2 3 4
Syllable
Garcia Binary rhythm and syllable weight in Portuguese 8 of 21
Methods
Duration in stimuli: 5σ
+ Durational pattern for secondary stress in stimuli
Recorded by native speaker with phonetic training
Initial
100ms
300ms
500ms
1 2 3 4 5
Syllable
Peninitial
100ms
300ms
500ms
1 2 3 4 5
Syllable
Garcia Binary rhythm and syllable weight in Portuguese 9 of 21
Main results
Initial stress overall favored over peninitial stress
4- and 5-syllable words: expected di�erence
Syllables: 4 (controls) Syllables: 5
HLHL LHHL LLHL HLLHL LHLHL LLLHL0%
25%
50%
75%
100%
Initi
al s
tres
s pr
efer
ence
LLLHL � HLLHL � LHLHL
Garcia Binary rhythm and syllable weight in Portuguese 10 of 21
Main results
Hierarchical logistic regression (50% and 95% CIs)
initial ∼ weight + (1 + weight | ID)
LL*
LH
HL
-1.5 -0.5 0.5 1.5
Posterior distribution (β)
I Positive β → initial secondary stress
Results interpreted relative to intercept (LL*)
Garcia Binary rhythm and syllable weight in Portuguese 11 of 21
Reaction time
Trends
Overall: faster RTs when choosing initial secondary stress
Syllables: 4 (controls) Syllables: 5
HLHL LHHL LLHL HLLHL LHLHL LLLHL
-0.1
0.0
0.1
0.2
RT
(st
d)
initialpeninitial
+ but even faster for 4-syl words
Speakers also more certain when choosing initial stress
Garcia Binary rhythm and syllable weight in Portuguese 12 of 21
Coda sonority
Trends
+ Initial stress preferred when coda contains liquid or nasal
I No clear trend in control items (4-syllable words)
Syllables: 4 (controls)Weight: HLHL
Syllables: 4 (controls)Weight: LHHL
Syllables: 5Weight: HLLHL
Syllables: 5Weight: LHLHL
obs LN obs LN obs LN obs LN0%
25%
50%
75%
100%
Sonority
Initi
al s
tres
s pr
efer
ence
Garcia Binary rhythm and syllable weight in Portuguese 13 of 21
Summary and discussion
I Overall: bias towards initial secondary stress
Stronger bias for 4σ words (clash avoidance)
I Location of Hσ seems to a�ect speakers’ judgements:
HLLHL � LHLHL
+ Lexical⇔ post-lexical planes (Dresher and van der Hulst 1998)
Why LLLHL � HLLHL?
Garcia Binary rhythm and syllable weight in Portuguese 14 of 21
Summary and discussion
50% ambiguity:
Speakers consistently find both options good
Speakers consistently find both options not so good
Syllables: 4 (controls) Syllables: 5
HLHL LHHL LLHL HLLHL LHLHL LLLHL0%
25%
50%
75%
100%
Initi
al s
tres
s pr
efer
ence
Hypothesis: presence of 2 Hσ seen as less natural
+ Confound in stimuli given lexical distribution of weight
Garcia Binary rhythm and syllable weight in Portuguese 15 of 21
Summary and discussion
≈90% of all words in Portuguese have 0/1 Hσ (Garcia 2014)
+ Same distribution when we control for frequency: (Tang 2012)
Hs: 0 Hs: 1 Hs: 2 Hs: 3
0 5 10 0 5 10 0 5 10 0 5 100
500
1000
1500
Frequency (log)
(Whether or not we examine 4-5σ words vs. all words)
Garcia Binary rhythm and syllable weight in Portuguese 16 of 21
Summary and discussion
Could duration in the stimuli explain speakers’ preferences?
A Initial:Peninitial ratio in ithitem with initial stress
B Initial:Peninitial ratio in ithitem with peninitial stress
+ Overall ratio A:B shows longer duration for initial σ
Main results (again)
0%
25%
50%
75%
100%
HLLHL LHLHL LLLHL
Initi
al s
tres
s pr
efer
ence
Ratios
1.5
1.6
1.7
1.8
1.9
2.0
HLLHL LHLHL LLLHL
Initial:Peninitial
Garcia Binary rhythm and syllable weight in Portuguese 17 of 21
Final remarks and future directions
Next steps:
A. Do we observe the same patterns with LLL]ω stimuli?
B. Can durational ratios alone explain LLLHL > HLLHL?
C. Does production mirror judgements?
Garcia Binary rhythm and syllable weight in Portuguese 18 of 21
References
Anttila, A. (2010). Word stress in Finnish. Handout, Yale University.
Boersma, P. and Weenink, D. (2020). Praat: doing phonetics by computer [Computer program].
Collischonn, G. (1994). Acento secundario em portugues. Letras de Hoje, 29(4):43–55.
Dresher, B. E. and van der Hulst, H. (1998). Head-dependent asymmetries in phonology: complexity and visibility. Phonology,
15(3):317–352.
Fernandes-Svartman, F., Abaurre, M. B., and Gonzalez-Lopez, V. A. (2008). Acento secundarop e intensidade em portugues
brasileiro. In Anais do 80◦ Encontro do CELSUL, pages 1–18, Porto Alegre.
Garcia, G. D. (2014). Portuguese Stress Lexicon. Comprehensive list of non-verbs in Portuguese. Available at
http://guilhermegarcia.github.io/psl.html.
Garcia, G. D. (2017). Weight gradience and stress in Portuguese. Phonology, 34(1):41–79.
Gordon, M. (2006). Syllable weight: phonetics, phonology, typology. Routledge, New York.
Hualde, J. I. and Nadeau, M. (2014). Rhetorical stress in Spanish. In van der Hulst, H., editor, Word stress: theoretical andtypological issues, pages 228–254. Cambridge University Press, Cambridge.
Lehiste, I. (1965). The function of quantity in Finnish and Estonian. Language, 41(3):447–456.
Major, R. C. (1985). Stress and rhythm in Brazilian Portuguese. Language, 61(2):259–282.
Moraes, J. a. (2003). Secondary stress in Brazilian Portuguese: perceptual and acoustic evidence. In XVth ICPhS, pages
2063–2066, Barcelona.
Pater, J. (2000). Nonuniformity in English secondary stress: The role of ranked and lexically specific constraints. Phonology,
17(2):237–274.
Tang, K. (2012). A 61 million word corpus of Brazilian Portuguese film subtitles as a resource for linguistic research. UCLWorking Papers in Linguistics 24: 208–214.
Wetzels, W. L. (2007). Primary word stress in Brazilian Portuguese and the weight parameter. Journal of PortugueseLinguistics, 5:9–58.
Garcia Binary rhythm and syllable weight in Portuguese References
Thank you!
This research is funded by Ball State University [19-0214]
Garcia Binary rhythm and syllable weight in Portuguese Appendix
Durational di�erences
Iniial vs. peninitial
HLLHL LHLHL LLLHL
initialpeninitial
1 2 3 4 5 1 2 3 4 5 1 2 3 4 5
100ms
300ms
500ms
100ms
300ms
500ms
Syllable
Garcia Binary rhythm and syllable weight in Portuguese Appendix